Laboratory validation of confirmatory tests for rabies diagnosis: Approaches to reduce animal use and facilitate sample collection.

Rabies is an encephalitis caused by rabies virus, whose transmission occurs upon contact with infected animals' saliva. The diagnosis is usually performed post-mortem through a direct fluorescent antibody test (DFAT). If the DFAT results are negative, they must be confirmed with an isolation test, usually the mouse inoculation test (MIT), which implies the suffering and death of the animals, high costs and most importantly, up to 28 days to confirm a negative result. Another issue related to rabies diagnosis is the sample collection and storage, which is critical for the rabies virus' RNA genome. Thus, this study aimed to evaluate (i) reverse transcriptase polymerase chain reaction (RT-PCR) and Rabies Tissue Culture Infection Tests (RTCIT) in comparison to DFAT and MIT and (ii) FTA ® cards as an alternative sample collection and preservation method. Eighty animal samples were evaluated through DFAT, RTCIT and RT-PCR; MIT was performed only in DFAT-negative samples. FTA ® cards were evaluated with a subset of 64 samples, with sufficient material for imprinting. Sensitivity, specificity, positive (PPV) and negative predictive values (NPV), agreement and Cohen's kappa were calculated for each test combination. RTCIT had higher sensitivity (92.5%) and RT-PCR had higher specificity (92.3%) compared to DFAT. The combination of tests enhanced sensitivity, NPV and Cohen's kappa (considering positive results by RTCIT or RT-PCR), and specificity and PPV (when both tests were concordant). The PCR based on FTA ® cards as sample source was specific (84.6%-96.2%) but presented lower sensitivity (29.7%-73.0%), although it could detect as positive four DFAT-negative samples. RTCIT and RT-PCR may be used as confirmatory tests in DFAT-negative samples. Moreover, FTA ® cards may be helpful for sample collection in field situations where a long time is needed until the sample undergoes laboratory testing.